Bot Balance next steps (as of 1.012)

[keith.lamothe]

For the Doombots, I didn't have a save of that exact mission, no.

So I went and tried generating one over and over to see if I couldnt get a similar situation.

I didn't get quite the silly number of them, but I did find one that still has.... alot of them, way more than it should, AND then some added special bots too.   I went and snuck through this one just to light everything up, so should be easy to see how many there are.

Thanks for the save.  I took a look and actually the distribution is looking pretty much as expected:

BotType	Category	CountPresent	PITAR	ExpectedPercentageOfCategory	ActualPercentageOfCategory
TreadBot	NormalEnemy	23	12	0.377	0.371
NinjaBot	NormalEnemy	13	16	0.283	0.21
DoomBot	NormalEnemy	7	40	0.113	0.113
EMPBot	NormalEnemy	10	40	0.113	0.161
SilenceBot	NormalEnemy	9	40	0.113	0.145
BodyBot	SpecialSeedingEnemy	1	4	0.426	0.167
MantaBot	SpecialSeedingEnemy	2	4	0.426	0.333
GalaxyBot	SpecialSeedingEnemy	1	16	0.106	0.167
RadarBot	SpecialSeedingEnemy	2	40	0.043	0.333

The Expected will generally not be exactly the Actual (particularly with special seeding bots, but DoomBot's not one of those) since there's some randomization after the frequencies are established, but in this case the DoomBot seeded exactly the expected number: 7.

That said, if the mission had had a normal bot type with a PITAR of 4 or 5, you would have seen fewer DoomBots.  In other words: the lack of a true "fodder" bot type in the random selection made all 5 normal types that were picked considerably more frequent than they would have been.  One of the changes I'm planning for the next version is to make sure each mission picks at least 1 "fodder" type (PITAR in the 4-7 range, possibly) to make sure this kind of thing doesn't happen this bad.

[nas1m]

For the Doombots, I didn't have a save of that exact mission, no.

So I went and tried generating one over and over to see if I couldnt get a similar situation.

I didn't get quite the silly number of them, but I did find one that still has.... alot of them, way more than it should, AND then some added special bots too.   I went and snuck through this one just to light everything up, so should be easy to see how many there are.

Thanks for the save.  I took a look and actually the distribution is looking pretty much as expected:

BotType	Category	CountPresent	PITAR	ExpectedPercentageOfCategory	ActualPercentageOfCategory
TreadBot	NormalEnemy	23	12	0.377	0.371
NinjaBot	NormalEnemy	13	16	0.283	0.21
DoomBot	NormalEnemy	7	40	0.113	0.113
EMPBot	NormalEnemy	10	40	0.113	0.161
SilenceBot	NormalEnemy	9	40	0.113	0.145
BodyBot	SpecialSeedingEnemy	1	4	0.426	0.167
MantaBot	SpecialSeedingEnemy	2	4	0.426	0.333
GalaxyBot	SpecialSeedingEnemy	1	16	0.106	0.167
RadarBot	SpecialSeedingEnemy	2	40	0.043	0.333

The Expected will generally not be exactly the Actual (particularly with special seeding bots, but DoomBot's not one of those) since there's some randomization after the frequencies are established, but in this case the DoomBot seeded exactly the expected number: 7.

That said, if the mission had had a normal bot type with a PITAR of 4 or 5, you would have seen fewer DoomBots.  In other words: the lack of a true "fodder" bot type in the random selection made all 5 normal types that were picked considerably more frequent than they would have been.  One of the changes I'm planning for the next version is to make sure each mission picks at least 1 "fodder" type (PITAR in the 4-7 range, possibly) to make sure this kind of thing doesn't happen this bad.

Sounds good!

[nas1m]

There might have to be done something on bosses as well.
Especially in terms of the number of level-ups they get granted and in terms of how many bosses are pulled into an Assassination mission.

In my current game of Expert what I got is this:





For perspective, I am on day 33, I think. How am I supposed to kill these guys - especially all at once (ie. killing these step by step over multiple HQ defense missions might be another story). Currently there is not remotely any way for my Exos to deal as much damage as I would need during a single mission - even more so if taking boss ranges into account...

An alternative could be to allow the player to "opt-out" out of an Assassination mission after at least one boss has been killed (similar to what we discussed - but not yet agreed on - for Factory missions. Currently these might be a tad too tough ...
EDIT: Mantised here.

[keith.lamothe]

FYI, I'm targeting getting Bionic 1.012 out Saturday noon-ish eastern time.

On the assassination mission, it looks like the main problem is that it seriously pumped up that one type of boss before its time.  I'm making some changes to that logic that should probably help a lot there, so we'll see how it goes.

I think the quantity of bosses pulled into an assassination mission is set to exactly 3, though I might be remembering wrong.

[nas1m]

FYI, I'm targeting getting Bionic 1.012 out Saturday noon-ish eastern time.

On the assassination mission, it looks like the main problem is that it seriously pumped up that one type of boss before its time.  I'm making some changes to that logic that should probably help a lot there, so we'll see how it goes.

I think the quantity of bosses pulled into an assassination mission is set to exactly 3, though I might be remembering wrong.

Well, it's 10 in the mission the screenshots come from.
Something might be awry here .

OT: And thanks for the custom map tile enum stuff! You are the best - I was just about to ask you for them .
Oh, and you might want to mark this one as resolved as well!

[keith.lamothe]

Well, it's 10 in the mission the screenshots come from.

10 bosses in one assassination mission?  Eep.  Do you have a save of that mission?

[nas1m]

As it happens ...
(Well, there are only 9 left but I guess this counts as well)

EDIT: Oh, and the changes you made for 1.012 look like they are shaping up nicely. Should definitely help to improve the level-up distribution situation. And great call on the predetermined bot mix as well. I think this is one of the major reasons why people only ever trying the "normal" difficulties could feel like the game is samey and or boring...

[Misery]

Well, it's 10 in the mission the screenshots come from.

10 bosses in one assassination mission?  Eep.  Do you have a save of that mission?

Yeah, I've seen 10 to 12 before.  Pretty frequently actually.

I'll include a screenshot here.  11 bosses in there, I believe.

Dropping it too far though might drop the challenge of this mission by too much, though.

I'll note that I did, in fact, beat that mission.  No Exo losses

[keith.lamothe]

Dropping it too far though might drop the challenge of this mission by too much, though.

True.  I guess mainly I want it to no longer be so random, as currently it can literally fail to seed any bosses (rarely) or seed a boss on every single spot if it has enough (also rarely).

Possibly make it 10 on Misery, 8 on Expert, etc?

[nas1m]

Dropping it too far though might drop the challenge of this mission by too much, though.

True.  I guess mainly I want it to no longer be so random, as currently it can literally fail to seed any bosses (rarely) or seed a boss on every single spot if it has enough (also rarely).

Possibly make it 10 on Misery, 8 on Expert, etc?

I would be more in favor of 5 for expert and 8 for Misery tbh...

[Misery]

Either is fine by me.

[nas1m]

Keith, could you provide a table of the boss bot stats as well?
I have a feeling that TethysBot might be in for nerf ...

[ScrObot]

I just finished a game on Hard, started in half in 1.011, half in 1.012. Bot balance seems pretty good, I didn't come across anything I felt was stupidly unfair. I ended up losing 2 missions, once to a BodyBot, the other in an Assassination mission (TethysBots and DragonBots, I believe). I don't have any specific complaints regarding the recent changes as of this moment, so... success?

BTW, pretty much the only way I could deal with the boss bots was to virus them and let them waste their ammo on each other.

[nas1m]

BTW, pretty much the only way I could deal with the boss bots was to virus them and let them waste their ammo on each other.

This is also true for my current game on expert. Thus my question for the boss bot stats above ...

[keith.lamothe]

Yes, there will be boss stats; I've just been very busy the last few days